Return to BSD News archive
Path: euryale.cc.adfa.oz.au!newshost.carno.net.au!harbinger.cc.monash.edu.au!munnari.OZ.AU!spool.mu.edu!howland.erols.net!news.mathworks.com!fu-berlin.de!news.belwue.de!rubin.noc.dfn.de!schweikh From: schweikh@rubin.noc.dfn.de (Jens Schweikhardt) Newsgroups: comp.unix.bsd.freebsd.misc Subject: Q: how to recompile libm Date: 12 Feb 1997 11:04:24 GMT Organization: Regional Network of Baden-Wuerttemberg, Stuttgart, FRG Lines: 29 Message-ID: <5ds83o$li4$1@news.belwue.de> NNTP-Posting-Host: rubin.noc.dfn.de Xref: euryale.cc.adfa.oz.au comp.unix.bsd.freebsd.misc:35344 hello, world\n In a recent thread I read that FreeBSD's libm is not compiled for floating point hardware (but rather emulates it). It was suggested to edit /etc/make.conf to include HAVE_FPU=YES and recompile libm. I cd to /usr/src/libm and for lack of READMEs describing the build process in detail I just typed make. Needless to say it didn't work :-) This may of course be due to my ignorance. The make stopped when compiling ieee/support.c due to prototype differences of some function (whose name I can't remember but I'll dig it up if you need it) with /usr/include/math.h. I'm probably too clueless here. Are there any make variables I have to set, like MACHINE or whatever to successfully build a new libm? Thanks for any hints. PS: All this is happening with 2.1.6 freshly off of the WC cdrom I got a week ago. Jens -- Jens Schweikhardt <schweikh@noc.dfn.de> SIGSIG -- signature too long (core dumped)